Advances in the clinical management of patients with non-small-cell lung cancer (NSCLC) have been slow, and the 5-year survival, which is used as a benchmark for putative cure from this disease, is only 14%. Reasons are that most patients present in advanced stages of the disease when response rates (25-35%) to chemotherapy are low and complete and prolonged partial remissions rare, despite the availability of several newer agents. Clearly an innovative and novel therapeutic strategy is needed to improve survival in this disease. Recent advances in the identification of molecular determinants of poor outcome from NSCLC hold the promise for therapeutic decisions based on individual patients' tumor's molecular profile. We hypothesize that we will improve response rates, median, and 1-year survival in patients with advanced NSCLC by treating with chemotherapy selected on the basis of the patient's tumor's molecular profile. Results generated by us and others suggest that the outcome of patients with lung cancer is closely associated with expression of the genes ERCC1 and RRMI. High ERCC1 and RRM1 gene expression predicts for resistance to platinum agents and gemcitabine, respectively. Using ERCC1 and RRM1 as predictive markers for chemotherapy response and survival we designed a phase II clinical study. The goal is to determine the response rates (RR) in newly diagnosed patients with advanced NSCLC who are treated with a chemotherapeutic regimen assigned to them on the basis of ERCC1 expression and RRM1 expression. Prior to treatment we will measure the level of ERCC1 and RRM1 expression in the patient's tumor, on the basis of which the patient will be assigned specific doublet chemotherapy. After treatment, response rates will be assessed as primary endpoint. Additional measures of treatment efficacy are progression free survival (PFS), median survival MS), and 1-yr survival(1-yr S) in patients with advanced NSCLC treated using this rational algorithmic approach. Patients will be followed till death and survival curves will be generated.